Equilibrium Polymerization of Sulfur - Journal of the American

E; Energy & Fuels · Environmental Science & Technology · Environmental Science & Technology Letters; I; Industrial & Engineering Chemistry · - Journal...
0 downloads 0 Views 250KB Size
ARTHUR V. TOBOLSKY AND ADI EISENBERG

780

[ C O N T R I B U T I O N F R O M THE

FRICK C H E M I C A L LABORATORY, PRISCETON

VOl. 81 UNIVERSITY]

Equilibrium Polymerization of Sulfur B Y ARTHURv. TOBOLSKY 4 N D AD1 EISENBERG RECEIVED AUGUST21, 1958

A simple and unified theory is presented describing the ring-chain equilibrium over the entire liquid range of sulfur One single formula is given describing the number average degree of polymerization a t any temperature in that range, and the AH and A S values for the initiation and propagation reactions are given.

I n previous papers, 1--3 a theory describing equilibrium polymerization in the presence of an initiator was developed and successfully applied to the polymerization of ecaprolactam by water. The reversibility of the viscosity-temperature curve of sulfur4 indicates that here also an equilibrium polymerization is involved, differing, however, from caprolactam by the absence of an external initiator. The conditions a t equilibrium a t any temperature for the polymerization of sulfur may be derived as I

M,,-I*

Let N

=

-

+ M +M,*;

Mn*= K3‘“-’K‘Mm ( 4 )

total concentration of polymer molecules. m

N

M,*

= n=l

N = K’MI1

+ K3’M + (K3’M)’ + ( K 3 ’ ~ ’ d+) ~ . . . . . .I

Let W = total concentration of monomer segments ( S g units) incorporated in the polymer. m

It is obvious that W I N = P , where P number average chain length (in terms units).

We also know that ivro =

M

+ w = )If + (1 -K‘M K~’A’@)~

(1) A. V. Tobolsky, J . Polymer Sci., 26, 220 (1937) ( 2 ) A. V. Tobolsky, addendum t o above, in press.

(3) A. Eisenberg and A. V. Tobolsky, THISJ O U R N A L , 81, in press (1959). (4) R. F. Bacon and R. Fanelli, ibid.,66, 539 (1943).

and inserting equation 7 into equation 8, we get the useful calculational equation

from which the entire chain-length vs. temperature curve may be calculated by knowing only the parameters K’ and K,’ (which are temperature dependent, and subject to the usual laws governing equilibrium constants) and which equals 3.90 moles/kg. (We employ this concentration unit rather than moles/liter because of the temperature dependency of the density of sulfur and because of the density change when monomer is converted to polymer, which may also be temperature dependent, and for which no precise data seem to be available.) It must be emphasized that in deriving equation Sa, no assumptions were made which restrict the validity of the formula to any temperature region; it should be applicable in the entire liquid range. Furthermore, the derivation does not take into account the probable formation of rings larger than Ss, nor does i t depend on the detailed mechanism by which equilibrium is reached (this is shown for a similar case in reference 3. -4ppendix 11). The probable presence of Sa rings is also not taken into account, since their effect on the equilibrium is believed to be negligible. As has been stated, only a knowledge of K‘ and K,‘ for any temperature is required for the determination of P and M a t that temperature, and, conversely, if P and iM are known, K’ and K3‘ can be determined. Unfortunately, neither P nor M can be determined experimentally with high precision. Hammick, Cousins and Langfordj attempted to determine 114 by measuring the percentage of sulfur of small quenched droplets insoluble in solvent. This is perhaps the most direct estimate of the percentage of polymer in the liquid, yet i t is clearly not ideal, due, inter alia, to the difficulty of quenching the liquid quickly enough to obtain true equilibrium values. No experimental estimate of the molecular weight or the chain length appears to be available. The molecular complexity of liquid sulfur was investigated on a theoretical basis by G. Gee6in an important pioneering contribution. We shall use his theoretical results as primary data in the absence of comprehensive experimental data. Gee’s theories are necessarily quite cumbersome because he had to draw together very sparse experimental data and very ingenious assumptions to obtain a comprehensive semi-quantitative description of the sulfur equilibrium. His description does satis( 3 ) Hammick, Cousins and Langford, J . Chem. Soc., 797 (192‘3). ( 6 ) G . Gee, Trans. F a r a d a y Soc., 48, 515 (1952).

Feb. 20, 1959

781

EQUILIBRIUM POLYMERIZATION OF SULFUR

1.0

.2

a

t

L

t

.

I

\

lo-”

1

I5

20

25

-.

1

1

I

I

2.0

1.5

1000

Fig. l.-K’vs. 1/T.

Fig. 2.-KS’ vs.

factorily incorporate all the known experimental facts as far as these exist. On the other hand, an example of the complexity of Gee’s theory is that he had to use two quite distinct treatments below and above the “transition temperature” a t 432’K., neither of which is valid in the immediate vicinity of 432’K. The theory given in equations 1-8, on the other hand, does give a satisfactory explanation of all of Gee’s results on the basis of a single, simple treatment which is valid over the entire liquid range as far as we can determine. Gee’s results, which we use in the sense of primary experimental data, are given in equations 9 through 12

In P = In P m

5 -1 - - + -1I n + 2HR(T 2

& =

;m)

H4 -~ Hs - Hi

I

I

~

,

I

/

25

1/T.

4 -

b

- (10) &,

(11)

applicable above the “transition temperature” T+ (at which polymer seems to appear suddenly), and

I

-

I

applicable below TQ,where qi is the weight fraction of polymer in the liquid, H4 refers to the reaction Si* G Si-** f SS, Pm is the maximum chain length (taken as lo6 atoms), T , is the temperature a t which P, is reached and qim is the corresponding weight fraction of the polymer, HS refers to the reaction Si* Si.,* Sr*, and P, is the chain length a t the transition temperature. I n a later

+

I

-.1000 T

1

l

l

l

l

l

l

l

,

l

~

publication’ a final set of numerical values of the constants used in equations 9 through 12 were (7) F. Fairbrother, G. Gee and G. T. Merrall, J . Poly. Sci., 16, 459 (1955).

~

~

~

782

dRTllUR

5‘. TOBOLSKY AND ,4DI EISENBERG

I?(>l. 81

TABLE I

i’

I

c

-.I

‘EXPEQIMENTAL’ P O I N T S

I

I

- CALCULATED

b

I

1

i

_

,



4000 4503

Fig. 4.-P

T.

1

=

I

O K .

Pa

.If

h=l

Ka’

385 2.21 0,1402 4n0 3 . 3x ,1805 3 410 5 . r ~ ,2050 4 420 9.44 ,2293 5 425 16.4 ,2405 6 428 27.6 .2470 7 4.10 57.9 ,2525 8 440 112,300 3 65 5 . 3 2 X 10-12 ,2736 9 450 113,909 3 . 3 6 1.22 X 2970 10 460 94,500 3.14 2 . 7 1 X IO-” ,3185 11 470 75,800 2.89 6.09 X lo-” ,3460 12 490 46,009 2.52 2.59 X ,3976 13 510 28,400 2.21 9 . 4 5 X 10-1° ,4514 14 540 13.870 1 . 8 6 5 . 6 9 X 10-B ,5366 15 580 5,750 1 . 5 2 2 . 7 1 X 10-8 ,6527 “ T h e number average chain length as obtained from formula 10 or 12 refers t o the number of S atoms in the chain. In this table, as also throughout this article, P is defined as the number average chain length in terms of Ss units. Therefore, the values for P obtained from Gee’s formulas were divided by 8 to yield the values given in Table 1. 1 2

Plots of log K’ vs. 1/T and log K3‘ vs. 1/T obtained from Table I are shown in Figs. 1 and 2. Straight lines are obtained over most of the “experimental” range taken from Gee’s computations. We assume that the correct values of K‘ and K,‘ follow the van’t Hoff law over the entire liquid range. The equations for the straight lines shown in Figs. 1 and 2 are

-

500 0

tu. T.

K ’ = 1.137 X IOK exp[-

given

Hd

No.

Temp

- 3200 cal./mole

T+ = 432’K.

y];

cal deg. mol cal AH’ = 32,800-2mole A S ’ = 23.0-

H6 = 24,000-35,000 cal./mole (we used 30,000 cal./mole)

Ks’ = 10 43 exp[

which yield

- ;]”->:A

6, = 0.0964 T, = 442 8°K.

( T , and $, are related by equation 9). Using the above constants in equations 9 through 12, expressions for P and M (where M = Mo(l +) = 3.90(1 - 4 ) ) can be obtained as explicit functions of temperature. Once P and 11.1 are known a t a given temperature, K‘ and KY’ can be calculated by equations 7 and 8. A list of numerical values of P and 11.1 obtained from Gce’s formulas and the corresponding values of K’ and K,’ are given in Table I. I t is of course not surprising that one can compute values of K’ and K3’ from P and M a t a single temperature. The check on the validity and usefulness of our approach is that the values of K‘ and K3’ a t different temperatures should follow van’t Hoff’s law, and give sensible values for AH’ and AH,,‘.

-

cal. deg. mole cal. AHa’ = 3,170mole A&’ = 4.63 ___

where the entropies and enthalpies refer to a standard state of 1 mole,’kg. Using equations 13 and 14 as absolutely valid and substituting numerical values of K’ and K3’ in equations 7 and 8, theoretical values of P and M can be obtained a t all temperatures. These are plotted in Figs. 3 and 4 and compared with Gee’s values. The comparison shows excellent agreement. In summary, we have completely predicted the sulfur equilibrium curve by a simple and unified theoretical treatment, completely comprised in equations 7, 8, 13 and 14. Furthermore, equations 13 and 14 are merely the simplest expression of van’t Hoff’s law, with a non-temperature dependent AHoand ASo. PRINCETOS, NE\$‘JERSEY